This study forecasts the future cost of direct economic damages caused by natural disasters in Korea by using panel data regression. The authors first develop a balanced panel data spanning 2001-2012 for all 16 metropolitan cities and provinces in Korea. The result shows that a 1% increase in annual precipitation and impervious surfaces increases damage costs by 4.52 and 1.74%, respectively. In addition, the financial independence of local governments is negatively correlated with damage costs. The maximum annual damage costs from natural disasters through 2060 are estimated to be US$20.9 billion, which would be 1.03% of future Korean gross domestic product (GDP). Among the regions, Gangwon-do, Jeollabuk-do, and Jeollanam-do, which have a high percentage of impervious surfaces and a low financial independence rate, are expected to be the most vulnerable to natural disasters. Their estimated maximum annual damage costs are above US $5 billion by 2060, which would exceed 7% of their gross regional domestic product (GRDP). This study is the first attempt to estimate and forecast the damage costs from natural disasters using region-specific weather data in Korea. It suggests a need for a well-designed natural disaster management plan at both the central and local government levels.